Oil production engineers are daily faced with vast amounts of real-time sensory data, which is processed to extract important information about the state of the production system. Data processing and analysis is done by the human brain, with support from machines. Traditionally, valuable information about the uncertainty in the data is discarded during processing. This leads to estimates without uncertainty measures, which in some cases become meaningless.
Our paper and corresponding presentation is a part of the technical session named "Real Time Production Optimisation". We will present a data-driven approach to production estimation and optimization that carries uncertainty information all the way through the data pipeline. This ultimately results in better decision making and increased production.